基于热释电无线传感网络的人体目标识别系统

发布时间：2018-12-24 07:58

【摘要】：目标检测是人体运动识别的一个重要组成部分，其意义主要是识别和跟踪以人为主的运动行为目标。人体运动识别技术是高级人机接口、人体运动分析和智能监控等领域共同涉及的关键性技术。热释电红外传感器可敏锐感知到特定波长的红外辐射，，该物理特性使其适用于人体目标识别领域。本文主要研究如何利用红外传感器设计一种基于无线传感器网络的人体目标识别系统。论文的主要研究内容和方法可以归纳为以下四个方面：首先，介绍热释电红外传感器的结构、原理和特性，以及原有的基于热释电红外传感器的人体目标识别方法，并且提出了新的人体识别方法，将三个热释电红外传感器节点放置在正三角形的三个顶点，根据3个节点的相交感应区域将全部感应区分割为多个子区域，通过采集节点的人体目标的数据信息。其次，介绍Zigbee无线技术，包含技术参数、网络体系以及网络拓扑结构。根据Zigbee无线技术特性，设计监控终端，实现无线网络的传输。再次，重点设计了人体目标识别系统的硬件、软件两大部分。人体目标识别系统的硬件部分至少需要三个传感器节点，每个传感器节点由母板、传感器模块和Zigbee模块共同组成；软件部分设计了协议网关和监控终端，通过其实现对数据传输和监控。最后，通过BP神经网络对经由热释电传感器采集到的人体运动信息进行分析处理，并计算出人体目标识别系统的识别率。
[Abstract]:Target detection is an important part of human motion recognition. Human motion recognition is a key technology in advanced man-machine interface, human motion analysis and intelligent monitoring. Pyroelectric infrared sensor is sensitive to infrared radiation at specific wavelengths, which makes it suitable for human target recognition. This paper mainly studies how to use infrared sensor to design a human body target recognition system based on wireless sensor network. The main research contents and methods can be summarized as follows: firstly, the structure, principle and characteristics of pyroelectric infrared sensor are introduced, and the original human body target recognition method based on pyroelectric infrared sensor is introduced. A new human body recognition method is proposed in which the three pyroelectric infrared sensor nodes are placed at the three vertices of the triangle, and all the sensing regions are divided into multiple sub-regions according to the three nodes' intersecting induction regions. The data of the human body target of the node are collected. Secondly, Zigbee wireless technology is introduced, including technical parameters, network architecture and network topology. According to the characteristics of Zigbee wireless technology, the monitoring terminal is designed to realize the transmission of wireless network. Thirdly, the hardware and software of human target recognition system are designed. At least three sensor nodes are needed in the hardware part of human target recognition system. Each sensor node is composed of motherboard, sensor module and Zigbee module. In the software part, the protocol gateway and monitor terminal are designed, through which the data transmission and monitoring are realized. Finally, the human motion information collected by pyroelectric sensor is analyzed and processed by BP neural network, and the recognition rate of human target recognition system is calculated.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN92;TP212.9

【参考文献】